KING AIIn Development · v1.2.0 foundation

Controlled Execution

Let AI do useful work without handing it unrestricted authority.

01

Allow · Ask · Deny

KINGAIBOT binds approval to the task, tool and canonical argument hash. Dangerous actions fail closed.

02

Durable work

Task state and restart recovery are part of the public v1.2.0 engineering foundation.

03

Safer tools

Filesystem roots, HTTPS allowlists, SSRF protections and restricted shell boundaries reduce execution risk.

04

Audit and rollback

Hash-chained events and a verified update path support review and recovery. This is a development foundation, not a claim of zero residual risk.

05

Integration boundaries

Connect models, tools, data and identity separately, granting each permission only within the task scope.

06

Status and adoption

Confirm release status, known limits and deployment requirements before evaluation, custom work or production use.

Does this resemble your work?

Start with one real workflow.

Bring the problem, current systems and desired result. We will help identify a sensible first step.